Abstract
Background
Little is known about how human immunodeficiency virus (HIV) affects walking biomechanics, or about associations between HIV-related gait deviations, functional performance, and self-reported outcomes. This paper reports on (1) gait biomechanics and variability in people with HIV (PWH) and (2) associations with clinical tests, self-reported function, and falls.
Methods
A cross-sectional study tested consecutively sampled PWH (n = 50) and HIV-seronegative participants ([SNP] n = 50). Participants underwent 3-dimensional gait analysis, performed clinical tests (short walk and single leg stance tests with and without dual tasking, chair-rise tests, and a physical performance battery), and completed questionnaires about function and falls. Between-group comparisons were done using analysis of covariance. Linear correlations between gait variability, clinical tests, and patient-reported outcomes were established.
Results
People with HIV and SNP had comparable median ages (PWH = 36.6, interquartile range [IQR] = 32.0–45.6]; SNP = 31.1, IQR = 23.2–45.1). Compared with SNP, PWH walked slower (adjusted mean difference [MD] = −0.2 meters per second [m/s], 95% confidence interval [CI] = −0.3 to −0.1) with greater variability (adjusted MD = 14.7 variability score points, 95% CI = 9.9–19.5). Moreover, PWH were slower in five-times sit-to-stand (5STS) performance (adjusted MD = 1.9 seconds, 95% CI = 1.00–2.9). Significant deviations in hip kinematics (increased flexion; adjusted MDs = 2.4°–2.8°, P = .012–.016) and knee kinematics (reduced flexion; adjusted MDs = 2.3°–3.7°, P = .007–.027) were found in PWH during dual-task (DT) walking. The PWH’s 5STS moderately correlated with larger gait variability (usual pace r = −0.5; dual task r = −0.6), poorer self-reported mobility (r = 0.4) and self-care function (r = 0.5), and fear of falling (P = .003).
Conclusions
People with HIV presented with biomechanical deviations suggestive of a slowed and variable gait, especially under cognitive challenges. Five-times STS may be useful to screen for gait deviations in PWH.
Keywords: chair rise time, gait variability, HIV infection, kinematics, physical function
In a South African sample, relatively young adults with HIV demonstrated biomechanical walking deviations suggesting a slowed, unstable gait pattern, especially under cognitive challenges. Five-times sit-to-stand may provide a screening test for early detection of mobility deviations in clinical settings.
There has been growing recognition of human immunodeficiency virus (HIV)-associated morbidity, including declining function, impaired lower limb performance, and gait deviations [1]. A recent systematic review concluded that middle-aged PWH can present with gait impairments resembling those of much older people with previous falls [2]. Neuromotor impairments in PWH can result from many factors, including persistent low-grade inflammation, cellular senescence, and physiologic decline, which can accelerate aging and hasten geriatric syndromes such as falls by 10–15 years [3]. Lower limb impairment in PWH have been associated with poorer quality of life (QOL), reduced independence [3], future disability, and earlier mortality [4]. This can negatively affect health, work productivity, and healthcare utilization across the lifespan and will require long-term, potentially costly care unless early risk detection strategies can be developed.
Three-dimensional gait analysis (3DGA) sensitively detects changes in movement quality that may be subtle [5, 6]. There has been no research into how HIV affects gait biomechanics [2], or whether biomechanical deviations in PWH relate to declining function [7]. Because 3DGA can sensitively quantify the dynamic implications of impairments during a functional activity such as gait, it may be appropriate to detect early subtle changes in movement quality and lower limb function in PWH. However, interpreting discrete and interdependent gait data is complex. Composite gait scores calculated from biomechanical variables, such as the Enhanced Gait Variability Index (EGVI), account for interdependence and provide more user-friendly index scores that quantify overall gait pattern deviation [8]. The EGVI quantifies gait variability, which has been associated with unstable walking and future mobility disability [9, 10]. It is unfortunate that 3DGA is not routinely available in many poorly resourced healthcare settings. The clinical measures with which it correlates may therefore offer useful alternative clinical screening tests for early functional decline in PWH. This information will inform early interventions to minimize future morbidity and healthcare utilization.
This study reports novel information about gait impairments in PWH captured by 3DGA. It explores associations between 3DGA, self-reported function, falls-related outcomes, and pragmatic physical performance tests applicable to clinical practice, with the aim of identifying a sensitive clinical screening test for early functional decline in PWH.
METHODS
Study Context
This paper reports on a cross-sectional substudy nested within a longitudinal cohort (Cape Winelands HAART to HEART Study).
Ethics Approval and Patient Consent
The study protocol was approved by the Stellenbosch University Health Research Ethics Committee (N15/05/043) and the Western Cape Department of Health (WC_2016RP10_878). All participants provided written informed consent.
Participants
The study compared PWH with HIV-seronegative participants (SNP). Two primary care community health centers with HIV clinics in the Cape Winelands, South Africa, provided participants. The study population was described previously in a related study using the same participants [11]. Research nurses or HIV-counselors consecutively recruited participants between June 2016 and December 2017. Participants included those (1) aged 18–65, (2) not overweight (body mass index [BMI] <25 kg/m2), (3) independently ambulant, and (4) able to consent. Exclusion criteria were (1) pregnancy, (2) acute infection/illness, (3) peripheral neuropathy, (4) major neurological conditions, (5) neuromusculoskeletal impairments/injury that affected gait, (6) visual impairment, and (7) alcohol intoxication. The eligibility criteria aimed to exclude factors substantially affecting usual gait or the quality of 3-dimensional (3D) data. We thus included a cohort without obvious predisposing factors to gait/balance impairments. For example, peripheral neuropathy and obesity may cause functional gait and balance impairments and have been associated with falls in PWH [12–15]. Increased awareness of gait deviations within a population without obvious risk factors could contribute to more regular screening, which may be useful for reducing walking-related risks or injury severity. Furthermore, subcutaneous tissue movement is known to influence 3D data collected using body-worn inertial motion capture systems; this is less likely with lower BMI ranges [16].
Human immunodeficiency virus status was confirmed via rapid blood test. Participants provided written informed consent, including specific consent for HIV testing. Pre- and posttest counseling was available as needed.
Sample Size
Sample size was based on gait biomechanics data from the first 30 substudy participants. Preliminary analysis identified the most variable gait angle as ankle dorsiflexion at initial contact. Using the pooled standard deviation (SD) for this angle (8.4), 5% significance and 80% power, a sample of at least 45/group was required to detect a 5-degree minimum clinically important difference [17]. To allow for missing data, a sample of n = 50 per group was sought. Post hoc analysis determined that this sample had 83% power to detect a moderate correlation (r = 0.4) between clinical tests and a composite gait score (EGVI).
Three-Dimensional Gait Analysis
A wireless, portable inertial motion capture (IMC) system (myoMOTION Research Pro; Noraxon USA Inc.) was used to capture walking biomechanics (28 lower limb kinematics and 15 time-space parameters [TSPs] [18], which may be affected in elderly or fall-prone gait [19–24]). Seven inertial measurement units (IMUs) were placed on body segments according to a rigid lower body model provided by the system. The system was recalibrated before each walking trial [18] (see Supplementary Material for further detail).
Composite Gait Score
The EGVI was used as a composite score to quantify gait pattern variability. The EGVI is based on 5 TSPs associated with functional outcomes and falls risk (step length, step time, stance time, single support time, and stride velocity) [25, 26] and quantifies the difference in variability between reference and patient groups [25]. A score of 100 indicates “normal” gait variability; higher and lower scores, respectively, indicate increased and decreased variability [25]. See the Supplementary Material for further detail.
Dual Tasking
Dual tasking was introduced after protocol amendment, thus data were unavailable on early participants. Participants were required to perform some 3DGA, single leg stance (SLS), and clinical walking trials while counting aloud backwards [27]—ie, demonstrate dual control of motor and cognitive functions (see Supplementary Material).
Physical Performance Tests
These included the 6-meter walk test (6mWT) [28], the 30-second SLS (eyes open and closed) [29], the 5-times sit-to-stand test (5STS), and the Health ABC Physical Performance Battery (PPB) [30]. The Supplementary Material provides details regarding these tests and their scoring. Because the 6mWT, 5STS, and SLS also form part of the PPB, these tests were scored in 2 ways: (1) as part of the PPB to obtain the overall battery score, and (2) as stand-alone clinical tests (eg, not converting the test times, speed, or repetitions to scores but using the raw values).
Self-Reported Outcomes (Function and Falls)
The EQ-5D-5L [31] is a standardized and concise health-related quality-of-life questionnaire that is suitable for use in PWH [32] and provides a profile of an individual’s function and a global health state [33] (see Supplementary Material for further information). Three function-related dimensions of the EQ-5D-5L measured self-perceived function (mobility, self-care, and usual activities) [33]. Fall-related history included self-reported falls in the past 12 months (any/none), number of falls (single versus ≥2), and fear of falling ([FOF] yes/no) [11].
Other Variables
Most demographic, health, and lifestyle information were obtained by the broader longitudinal study using questionnaires, medical folders, and interviews (sex, education, employment, income, medication, chronic comorbidities, HIV and antiretroviral therapy [ART] history, smoking, self-reported physical activity, and alcohol consumption). Other measures obtained from the broader study included anthropometry and laboratory results for CD4 count and viral load (dichotomized at detectable threshold [≥50 cp/mL]). Participants in the current substudy (1) completed a supplementary questionnaire to assess pain, cognition, depressive symptoms, and combination ART (cART) adherence (dichotomized as taking ART as prescribed all or most of the time, or not compliant) [34], (2) were screened for peripheral sensory neuropathy, and (3) underwent lower limb isometric strength testing [35]. The Supplementary Material provides further details.
Study Procedures
Data were collected at the 2 sites in quiet, hard-floored rooms. The study visit included questionnaire completion, physical measurements, 3DGA, and random-sequence functional test performance. To standardize the testing protocol and facilitate skin-attachment of IMUs, participants were barefoot and dressed comfortably. Practice trials for 3DGA were performed after the IMUs were positioned, but before IMC calibration. For the 3DGA trials, conducted using standardized procedures, participants walked along a 10-meter walkway under 3 randomized conditions: (1) self-selected usual pace; (2) as fast as possible without running; and (3) with a DT. Participants performed 3 trials per condition. Walking started 1 meter before a line on the floor and ended after crossing a second line. A gait trial was deemed successful if 3 strides per limb were achieved.
Data Management
Inertial measurement unit data were transmitted wirelessly to a recording laptop and exported to MATLAB software (R2017a; MathWorks) for processing. Three successful 3DGA gait trials per condition were analyzed per participant. Cyclical gait events (foot contact and foot-off) were determine using a built-in algorithm in the software. This segmented the data into gait cycles normalized in time to 101 data points at 1%-time intervals. After gap-filtering, outcomes were (1) determined using analysis scripts and visualizations and (2) exported to MS Excel for analysis. The EGVI calculations were based on 5 TSPs generated within the 3DGA software, custom-formatted for use within the Excel EGVI macro [25]. Raw data were obtained from the 3 gait trials per participant. Within each trial, EGVI calculation was conducted if at least 3 consecutive values for each alternative parameter were available [36]. The macro calculates an overall EGVI score as a composite score around a z-score of 100, using the absolute difference between successive values of the same series (trial and leg) [25] (see Supplementary Material for further details).
Statistical Analyses
SPSS V25.0 (IBM Corporation, Armonk, NY) and STATA V14.2 (StataCorp 2015, Stata Statistical Software: Release 14; StataCorp, College Station, TX) were used. Statistical significance was set at P < .05. Data normality and homogeneity of variance for analyses of covariance (ANCOVA) were confirmed. Data were summarized, as relevant, using mean and SD, median and interquartile range, and frequencies and percentages. Floor and ceiling effects of >15% were considered significant for those clinical tests with maximum and minimum cutoff scores [37]. For univariate analyses, differences between PWH and SNP were determined using independent t tests, Mann-Whitney U tests, χ 2 tests, or Fisher exact tests. Clinical tests and 3DGA data were compared between groups using factorial ANCOVA. For clinical tests, covariables were age, sex, leg length, anxiodepressive symptoms, smoking status, and physical activity. For 3DGA, covariables included dimensionless (gravity- and leg-length-normalized) gait speed, age, and sex. Separate ANCOVA models were created for each dependent variable, treating clinical tests as continuous variables in all analyses [28]. Where significant interactions were noted between HIV-serostatus and the factors/covariables included in the ANCOVA, an “HIV × covariable” term was included in the final model [38]. Nonsignificant interaction terms were serially removed from the model [38].
Correlations between clinical tests and EGVI scores, self-reported function, and fall history were evaluated using Spearman’s rank (rs) correlation coefficients and Pearson product-moment (r) coefficients. Correlations of 0.20 ≤ 0.39 were weak; correlations of 0.40 ≤ 0.59 were moderate; correlations of 0.60 ≤ 0.79 were strong; and correlations of 0.80 ≤ 1.00 were very strong [39]. Associations between (1) clinical tests and (2) each falls variable were assessed using independent t tests or Mann-Whitney U tests.
RESULTS
Of 186 participants screened, 106 were eligible, and after further exclusion, 50 PWH and 50 SNPs participated in all tests (Supplementary Figure 1). Dual-tasking data were collected for 41 of 50 PWH and 47 of 50 SNP. Due to insufficient detection of consecutive strides, usual-pace EGVI scores were calculated for 44 PWH and 29 SNP, DT EGVI for 26 PWH and 30 SNP, and fast EGVI for 0 participants.
Participant Characteristics
There were significant differences between PWH and SNP regarding age, sex, leg length, muscle strength, physical activity, smoking, polypharmacy, anxiety-depression, self-reported function, and falls history (Table 1).
Table 1.
Participant Characteristics
| Characteristic | PWH (n = 50) | SNP (n = 50) | P |
|---|---|---|---|
| Age, median (IQR) | 36.6 (32.0– 45.6) | 31.1 (23.2– 45.1) | .017 |
| Younger than 50, n (%) | 40/50 (80.0) | 43/50 (86.0) | .424 |
| Women, n (%) | 29/50 (58.0) | 40/50 (80.0) | .017 |
| Education < Grade 12, n (%) | 29/50 (58.0) | 28/50 (56.0) | .840 |
| Employed, n (%) | 29/50 (58.0) | 20/50 (40.0) | .072 |
| Total monthly household income, n (%) | .645 | ||
| <R1000 | 16/50 (32.0) | 12 (24.0) | - |
| R1000–R4999 | 22/50 (44.0) | 24 (48.0) | - |
| R5000–R9999 | 8/50 (16.0) | 7/50 (14.0) | - |
| R10 000–R20 000 | 4/50 (8.0) | 5/50 (10.0) | - |
| >R20 000 | 0/50 (0.0) | 2/50 (4.0) | - |
| Height in meters, mean (SD) | 1.6 (0.1) | 1.6 (0.1) | .061 |
| BMI, kg/m2, mean (SD) | 21.5 (4.8) | 20.4 (5.0) | .677 |
| Leg length, cm, mean (SD) | 86.8 (5.5) | 84.5 (4.1) | .016 |
| Muscle strength, Newton | |||
| Plantarflexors, mean (SD) | 169.2 (36.2) | 181.9 (30.3) | .059 |
| Dorsiflexors, mean (SD) | 133.0 (23.9) | 131.8 (23.6) | .798 |
| Knee extensors, mean (SD) | 189.1 (38.0) | 203.8 (34.3) | .045 |
| Knee flexors, mean (SD) | 147.1 (32.6) | 161.1 (26.6) | .020 |
| Hip extensors, mean (SD) | 216.9 (39.0) | 212.2 (34.3) | .523 |
| Hip flexors, mean (SD) | 205.5 (37.3) | 193.8 (28.2) | .078 |
| Hip abductors, median (IQR) | 103.6 (86.4–126.4) | 109.2 (100.1–125.8) | .082 |
| Physical activity, n (%)a | .033 | ||
| None | 6/50 (12.0) | 15/50 (30.0) | - |
| Low | 22/50 (44.0) | 23/50 (46.0) | - |
| Moderate to high | 22/50 (44.0) | 12/50 (24.0) | - |
| Alcohol use over past 12 months, n (%) | .334 | ||
| None | 24/50 (48.0) | 19/50 (38.0) | - |
| Light | 18/50 (36.0) | 27/50 (54.0) | - |
| Moderate | 6/50 (12.0) | 3/50 (6.0) | - |
| Heavy | 2/50 (4.0) | 1/50 (2.0) | - |
| Current smoker, n (%) | 26/50 (52.0) | 36/50 (72.0) | .039 |
| Injection drug use, ever, n (%) | 1/50 (2.0) | 0/50 (0.0) | .500 |
| Nonantiretroviral polypharmacy, n (%) | 16/50 (32.0) | 5/50 (10.0) | <.001 |
| Multimorbidity, n (%) | 4/50 (8.0) | 2/50 (4.0) | .678 |
| Anxio-depressive symptoms, EQ-5D-5L, n (%) | 12/48 (25.0) | 2/50 (4.0) | .003 |
| Pain, MOS-HIV score, median (IQR) | 77.8 (55.6–88.9) | 88.9 (77.8–100.0) | .127 |
| Cognitive function, MOS-HIV score, median (IQR) | 80.0 (60.0–95.0) | 87.5 (75.0–95.0) | .076 |
| Self-rep. function: EQ-5D-5L Mobility problems, n (%) | 11/49 (22.5) | 4/50 (8.0) | .046 |
| Self-rep. function: EQ-5D-5L Self-care problems, n (%) | 6/49 (12.2) | 1/50 (2.0) | .047 |
| Self-rep. function: EQ-5D-5L Usual activities, n (%) | 10/49 (20.4) | 4/50 (8.0) | .077 |
| Any fall during past 12 months, n (%) | 17/50 (34.0) | 8/50 (16.0) | .038 |
| Number of falls during past 12 months, n (%) | .114 | ||
| Single | 10/50 (20.0) | 5/50 (10.0) | |
| Recurrent (2 or more) | 7/50 (14.0) | 3/50 (6.0) | |
| Fear of falling, n (%) | 10/50 (20.0) | 7/50 (14.0) | .425 |
| Years since HIV diagnosis, n (%) | |||
| <2 | 9/50 (18.0) | - | - |
| 2<5 | 16/50 (32.0) | - | - |
| 5–15 | 22/50 (44.0) | - | - |
| >15 | 3/50 (6.0) | - | - |
| Current CD4+ count, cells/μL, mean (SD) | 448.8 (233.0) | - | - |
| Detectable viral load, ≥50 cp/mL, n (%) | 25/46 (54.4) | - | - |
| On cART, n (%) | 45/50 (90.0) | - | - |
| First-line cART, n (%) | 38/45 (84.4) | - | - |
| Second-line cART, n (%) | 7/45 (15.6) | - | - |
| cART duration in weeks, median (IQR) | 119 (62–312) | - | - |
| cART regime, n (%) | |||
| PI-based | 9/45 (20.0) | - | - |
| NNRTI-based | 36/45 (80.0) | - | - |
| INI-based | 0/45 (0.0) | - | - |
| cART adherent, n (%) | 35/45 (77.8) | - | - |
Abbreviations: BMI, body mass index; cART, combination antiretroviral therapy; HIV, human immunodeficiency virus; INI, integrase inhibitor; IQR, interquartile range; MOS-HIV, Medical Outcomes Study HIV Health Survey; NNRTI, nonnucleoside reverse-transcriptase inhibitor; PI, protease inhibitor; PWH, people with HIV; rep., report; SD, standard deviation: SNP, HIV-seronegative participants.
aPhysical activity (PA) was based on self-report. Engaging in PA for <60 minutes/week (or <2 days/week) was deemed low PA. Engaging in mild to moderate exercise for >60 minutes/week, or participating in ≥30 minutes of moderate (or higher intensity) exercise for ≥5 days/week (or ≥150 minutes/week) were grouped together as moderate-to-high levels of PA [51]. Bold formatting indicates statistical significance at 5%.
Enhanced Gait Variability Index
For usual-paced gait, EGVI scores were significantly higher for PWH than SNP (age-sex adjusted MD = 14.7, standard error [SE] = 2.4, 95% confidence interval [CI] = 9.9–19.5, P < .001). Under the DT condition, PWH obtained higher scores than SNP, but the difference was not statistically significant after age-sex adjustment (MD = 6.4, SE = 3.2, 95% CI = 0.1–12.8, P = .052) (Figure 1).
Figure 1.
Enhanced Gait Variability Index (EGVI) scores in human immunodeficiency virus (HIV)-negative participants (SNP) and people with HIV (PWH) for usual-paced (UP) and dual-task (DT) walking. Plots show mean scores (labeled diamonds) with 95% confidence intervals (capped vertical error bars) as well as median scores (horizontal lines), score range (uncapped vertical error bars), and interquartile range (boxes).
Three-Dimensional Gait Analysis Outcomes
Table 2 presents adjusted comparisons between PWH and SNP for TSP and joint angles (Supplementary Tables 1A and 1B present results per group and task-condition). For usual-paced gait, PWH walked slower with longer stance and step times. Human immunodeficiency virus-serostatus showed significant main effects only for gait speed (F = 15.8, P < .001). Significant interactions were noted between (1) HIV-serostatus and sex on stance time (F = 4.9, P = .029) and step time (F = 5.0, P = .029) and also between (2) HIV-serostatus and gait speed on stance time (F = 14.9, P < .001) and step time (F = 12.3, P = .001). For both stance time and step time, HIV effects appear larger in men and slower walkers. For fast gait, all differences were nonsignificant, except for step time (longer in PWH). Under DT conditions, gait speed was significantly slower in PWH, with longer stride length and step time. A significant interaction existed for HIV-serostatus and gait speed on step time (F = 22.1, P < .001); the effects of HIV being more apparent at slower speeds.
Table 2.
Differences in Biomechanical Outcomes for Usual-Paced (n PWH = 50, n SNP = 48), Fast (n PWH= 45, n SNP = 45) and DT (n PWH = 39, n SNP = 47) Walking
| Three-dimensional gait analysis outcome (time/space parameter or angle) | Usual-paced | Fast | Dual task | |||
|---|---|---|---|---|---|---|
| Adjusted MD (95% CI) | PWH vs SNP | Adjusted MD (95% CI) | PWH vs SNP | Adjusted MD (95% CI) | PWH vs SNP | |
| Step length, cm | 1.3 (−0.4 to 3.0) | ↑ | 1.1 (−0.7 to 2.9) | ↑ | 2.0 (−0.1 to 4.1) | ↑ |
| Normalized step lengtha | 0.0 (−0.0 to 0.0) | ↑ | 0.0 (−0.0 to 0.0) | ↑ | 0.0 (−0.0 to 0.1) | ↑ |
| Stride length, cm | 2.4 (−1.0 to 5.8) | ↑ | 2.1 (−1.5 to 5.7) | ↑ | 4.5 (0.3 to 8.8) | ↑* |
| Normalized stride lengtha | 0.0 (−0.0 to 0.1) | ↑ | 0.0 (0.0 to 0.1) | ↑ | 0.1 (−0.0 to 0.1) | ↑ |
| Cadence, steps/min | −2.1 (−4.6 to 0.5) | ↓ | −2.7 (−5.8 to 0.3) | ↓ | −3.8 (−7.6 to 0.0) | ↓ |
| Normalized cadencea | −0.5 (−1.2 to 0.2) | ↓ | −0.8 (−1.6 to 0.1) | ↓ | −1.1 (−2.2 to 0.1) | ↓ |
| Stance time, ms | 20.0 (0.0 to 30.0) | ↑* | 10.0 (−0.0 to 30.0) | ↑ | 20.0 (0.0 to 50.0) | ↑ |
| Step time, ms | 20.0 (2.0 to 30.0) | ↑* | 10.0 (0.0 to 20.0) | ↑* | 190.0 (330.0 to 350.0)b | ↑*b |
| Single support time, ms | 10.0 (−0.0 to 30.0) | ↑ | 10.0 (0.0 to 20.0) | ↑ | 20.0 (0.0 to 40.0) | ↑ |
| Double support time, ms | 0.0 (−10.0 to 20.0) | ↑ | 10.0 (−10.0 to 20.0) | ↑ | 10.0 (10.0 to 30.0) | ↑ |
| Stance percentage, %GC | −0.1 (−0.7 to 0.5) | ↓ | −0.1 (−0.6 to 0.5) | ↓ | −0.3 (−1.1 to 0.5) | ↓ |
| Single support percentage, %GC | 0.2 (−0.5 to 1.0) | ↑ | 0.1 (−0.6 to 0.7) | ↑ | −0.1 (0.8 to 0.7) | ↓ |
| Double support percentage, %GC | 0.3 (−1.1 to 1.8) | ↓ | −0.1 (−1.1 to 0.9) | ↓ | −0.3 (−1.7 to 1.2) | ↓ |
| Gait speed, m/sb | −0.2 (−0.3 to −0.1) | ↓* | −0.1 (−0.2 to 0.0) | ↓ | −0.2 (0.3 to −0.1) | ↓* |
| Normalized gait speedb | −0.1 (−0.1 to −0.0) | ↓* | −0.0 (−0.1 to 0.0) | ↓ | −0.1 (−0.1 to −0.0) | ↓* |
| Pelvis tilt ROM, degrees | −0.3 (−1.4 to 0.7) | ↓ | 0.3 (−0.8 to 1.5) | ↑ | 0.1 (−0.9 to 1.1) | = |
| Peak pelvis anterior tilt, degrees | −0.2 (−2.0 to 1.7) | ↓ | 0.0 (−1.7 to 1.8) | = | 0.7 (−1.1 to 2.5) | ↑ |
| Pelvis obliquity ROM, degrees | −0.8 (−1.9 to 0.3) | ↓ | 0.4 (−0.9 to 1.6) | ↑ | −0.5 (−1.7 to 0.7) | ↓ |
| Pelvis rotation ROM, degrees | 0.5 (−1.2 to 2.3) | ↑ | 1.6 (−0.2 to 3.4) | ↑ | 1.3 (−1.1 to 3.6) | ↑ |
| Pelvis rotation at initial contact, degrees | 0.2 (−1.2 to 1.5) | ↑ | 0.8 (−0.7 to 2.4) | ↑ | 1.8 (0.3 to 3.3) | ↑* |
| Hip flexion ROM, degrees | 0.9 (0.6 to 2.3) | ↑ | 0.4 (−1.1 to 1.9) | ↑ | 2.4 (0.5 to 4.4) | ↑* |
| Hip flexion ROM during loading response, degrees | 0.2 (−0.9 to 1.4) | ↑ | 0.4 (−0.5 to 1.4) | ↑ | −0.4 (−1.6 to 0.8) | ↓ |
| Hip flexion ROM pre-swing; initial swing, degrees | 0.2 (−1.0 to 1.4) | ↑ | −0.4 (−1.4 to 0.6) | ↓ | −0.0 (−1.2 to 1.1) | = |
| Hip flexion at initial contact, degrees | 0.6 (−1.5 to 2.7) | ↑ | 0.1 (−1.9 to 2.1) | = | 2.8 (0.6 to 4.9) | ↑* |
| Peak hip flexion during swing, degrees | 0.00 | = | −0.4 (−2.3 to 1.5) | ↓ | 1.7 (−0.3 to 3.8) | ↑ |
| Peak hip extension during stance, degrees | 1.1 (−0.6 to 2.9) | ↑ | 1.1 (−1.4 to 3.6) | ↑ | 0.6 (−1.5 to 2.7) | ↑ |
| Hip abduction ROM during mid- stance, degrees | −0.2 (−1.0 to 0.6) | ↓ | −0.2 (−1.1 to 0.8) | ↓ | 0.7 (−0.1 to 1.5) | ↑ |
| Hip adduction ROM during loading response, degrees | −0.5 (−1.4 to 0.4) | ↓ | −0.3 (−1.2 to 0.6) | ↓ | −0.4 (−1.4 to 0.7) | ↓ |
| Hip internal rotation ROM, degrees | −0.7 (−2.5 to 1.2) | ↓ | 0.1 (−1.7 to 1.9) | = | −0.0 (−2.3 to 2.3) | = |
| Knee flexion ROM, degrees | −0.6 (2.4 to 1.2) | ↓ | 0.4 (−1.3 to 2.0) | ↑ | −2.9 (−5.1 to −0.6) | ↓* |
| Knee flexion ROM during stance, degrees | −0.3 (−2.0 to 1.3) | ↓ | −0.7 (−2.3 to 0.8) | ↓ | 0.8 (1.1 to 2.6) | ↑ |
| Knee flexion ROM from stance; swing, degrees | 0.24 (−1.6 to 2.1) | ↑ | 0.4 (−1.3 to 2.2) | ↑ | −2.3 (−4.6 to 0.0) | ↓ |
| Knee extension ROM from mid-stance; terminal stance, degrees | 0.9 (−0.9 to 2.7) | ↑ | 0.3 (1.6 to 2.2) | 0.737↑ | 2.3 (0.3 to 4.4) | ↑* |
| Knee flexion at initial contact, degrees | 1.5 (−0.9 to 4.0) | ↑ | −0.1 (−2.6 to 2.4) | 0.941↓ | 3.7 (1.0 to 6.5) | ↑* |
| Peak knee flexion during stance, degrees | 0.2 (−2.2 to 2.7) | ↑ | −1.5 (−3.7 to 0.7) | 0.169↓ | 0.1 (−2.8 to 3.0) | = |
| Peak knee extension during stance, degrees | −1.2 (−3.2 to 0.9) | ↑ | 0.8 (−0.9 to 2.6) | 0.357↓ | −3.2 (−5.5 to −0.9) | ↓* |
| Peak knee flexion during swing, degrees | 0.4 (−1.9 to 2.7) | ↑ | −0.8 (−2.9 to 1.2) | 0.426↓ | 0.1 (−2.7 to 2.9) | = |
| Ankle dorsiflexion ROM during stance, degrees | −0.9 (−2.3 to 0.5) | ↓ | −0.30 (−1.7 to 1.1) | 0.664↓ | −1.1 (−2.8 to 0.50) | ↓ |
| Ankle dorsiflexion ROM during swing, degrees | −1.5 (−3.7 to 0.6) | ↓ | −1.4 (−3.6 to 0.9) | 0.238↓ | −1.4 (−4.1 to 1.4) | ↓ |
| Ankle plantarflexion ROM, heel rise; toe-off (push off), degrees | −2.0 (−4.2 to 0.2) | ↓ | −1.3 (−3.5 to 0.8) | 0.219↓ | −1.9 (−4.6 to 0.9) | ↓ |
| Ankle dorsiflexion at initial contact, degrees | −0.1 (−1.5 to 1.2) | ↑ | −0.4 (−1.7 to 1.0) | 0.569↑ | 0.8 (−0.8 to 2.5) | ↓ |
| Ankle plantarflexion at toe-off, degrees | −0.9 (−3.4 to 1.6) | ↓ | −0.2 (−2.7 to 2.3) | 0.874↓ | −0.8 (−3.9 to 2.2) | ↓ |
| Peak ankle plantarflexion, degrees | −1.5 (−4.0 to 1.0) | ↓ | −0.4 (−3.0 to 2.2) | 0.748↓ | −1.9 (−5.0 to 1.2) | ↓ |
Arrows indicate the directional trend of the difference between PWH and SNP; asterisks indicate statistical significance at 5%.
Abbreviations: GC, gait cycle; MD, mean difference; ms, milliseconds; m/s, meters per seconds; PWH, people with HIV; ROM, range of motion; s, seconds; SNP, HIV-seronegative participants.
aSince normalized parameters are already scaled to leg length, multivariable models included absolute gait speed.
bGait speed (absolute and normalized) was only adjusted for sex and age in multivariable models.
cStandard error (SE) and associated CI were estimated with robust HC3 SE due to a significant Levene’s test (P =.02).
For joint angles, significant differences existed only for DT walking. These differences occurred at the pelvis, hip, and knee at various points during the gait cycle. A significant interaction was noted between HIV-serostatus and gait speed on knee flexion range of motion (ROM). Under DT conditions, the effect of HIV on reducing knee flexion ROM depended on the walking speed and appeared larger at slower speeds.
Clinical Test Performance
Table 3 summarizes differences in clinical test performance between PWH and SNP (Supplementary Table 2 presents results per group). Significant interactions existed between HIV-serostatus and anxiodepressive symptoms for both PPB gait speed subscores (F = 6.2, P = .015 and F = 4.9, P = .029 for usual and narrow tests, respectively). This suggested that the effect of HIV on gait speed scores depended on anxiodepressive symptoms (larger effect in the presence of symptoms). The PPB and its components demonstrated no significant floor or ceiling effects in either group, except for the balance score (highest level assessed: 30s SLS Eyes Open [EO]). Both groups demonstrated high ceiling effects for DT SLS. There were no between-group differences for SLS EO as part of the PPB subscore, and given the high ceiling effects, this test was not analyzed further as a stand-alone functional measure. A significant interaction existed between HIV-serostatus and anxiodepressive symptoms on usual-paced walking speed without DT (F = 5.5, P = .021). This interaction was similar for the PPB walking subscores. A significant interaction existed between HIV-serostatus and age on 5STS time (F = 6.1, P = .016), suggesting that the detrimental effect of HIV-serostatus on 5STS increases with age.
Table 3.
Clinical Functional Test Performance Differences Between People With HIV and HIV-Negative Participantsa
| Clinical Test | Adjusted MD (95% CI) | PWH vs SNP | Interpretation |
|---|---|---|---|
| Physical Performance Battery Score, Range 0–1 | |||
| Total score | −0.32 (−0.48 to −0.15) | ↓* | Worse in PWH |
| Balance score | −0.95 (−1.94 to 0.04) | ↓ | Worse in PWH |
| Usual walk score | −0.09 (−0.14 to −0.05) | ↓* | Worse in PWH |
| Narrow walk score | −0.16 (−0.26 to −0.07) | ↓* | Worse in PWH |
| Chair rise score | −0.09 (−0.14 to −0.04) | ↓* | Worse in PWH |
| Single-Leg Standing Test Time, Seconds | |||
| Eyes closed | −6.0 (−9.9 to −2.1) | ↓* | Shorter (worse) in PWH |
| Dual taskb | −4.7 (−9.8 to 0.4) | ↓ | Shorter (worse) in PWH |
| 6 m Walk Test Speed, m/s | |||
| Usual-paced | −0.3 (−0.5 to −0.2) | ↓* | Slower (worse) in PWH |
| Dual taskb | −0.2 (−0.3 to −0.1) | ↓* | Slower (worse) in PWH |
| Chair Rise Tests | |||
| 5× Sit-to-stand time, seconds | 1.9 (1.0 to 2.9) | ↑* | Slower (worse) in PWH |
| 30 s Sit-to-stand, repetitions | −4.8 (−8.6 to −1.0) | ↓* | Less (worse) in PWH |
Abbreviations: HIV, human immunodeficiency virus; m, meter; MD, mean difference; m/s, meters per second; PWH, people with HIV; s, second; SNP, seronegative participants.
aArrows indicate the direction of the difference between PWH and SNP; asterisks/bold print indicate statistical significance at 5%.
bNo dual-task data collected for n = 3 SNP and n = 9 PWH.
Correlations of clinical tests with the EGVI and self-reported outcomes in PWH Table 4 reports correlations of the clinical tests with the EGVI, self-reported function, and falls history for PWH. The 5STS correlated most strongly with EGVI scores, also showing moderate correlations with mobility function and self-care problems. People with HIV with slower 5STS demonstrated higher EGVIs, as did those reporting functional problems (Supplementary Figure 2).
Table 4.
Correlations Between Clinical Performance Tests, Enhanced Gait Variability Index, Self-Reported Function, and Fall Number in PWH
| PWH | EGVI UP | EGVI DT | Mobility Problems | Self-Care Problems | Usual Activity Problems | Fall No. in Past Year |
|---|---|---|---|---|---|---|
| PPB total score | −0.4* | −0.6* | −0.5** | −0.4** | −0.4** | 0.0 |
| PPB balance score | 0.1 | 0.2 | −0.3* | −0.3* | −0.1 | −0.5** |
| PPB usual walk score | −0.3*a | −0.3a | −0.5** | −0.4** | −0.4** | 0.1 |
| PPB narrow walk score | −0.3a | −0.2a | −0.4** | −0.4** | −0.4** | 0.2 |
| PPB chair-rise score | −0.4** a | −0.6** a | −0.4** | −0.5** | −0.3* | −0.2 |
| 6mWT speed | −0.3*a | −0.3a | −0.5** | −0.4** | −0.4** | 0.1 |
| 6mWT DT speed | −0.3a | −0.5* a | −0.4* | −0.1 | −0.2 | −0.1 |
| 5STS time | 0.5** a | 0.6** a | 0.4** | 0.5** | 0.3* | 0.2 |
| 5STS pace (stands/second) | −0.5** a | −0.6** a | −0.4** | −0.5** | −0.3* | −0.2 |
| 30sSTS repetitions | −0.4** a | −0.5** a | −0.4* | −0.5** | −0.4** | −0.1 |
| 30sSTS pace (stands/second) | −0.5** a | −0.5** a | −0.4* | −0.5** | −0.4** | −0.1 |
| SLS EC time | −0.2 | −0.0 | −0.5** | −0.5** | −0.3* | −0.2 |
| SLS DT time | 0.1 | 0.2 | 0.0 | −0.1 | −0.3 | −0.2 |
Abbreviations: 30sSTS, 30-Second Sit-To-Stand Test; 5STS, Five-times Sit-To-Stand Test; 6mWT, Six-meter Walk Test; DT, dual-task; EC, Eyes Closed; EGVI, Enhanced Gait Variability Index; PWH, people with HIV; PPB, Health ABC Physical Performance Battery; SLS, Single Leg Stance Test; UP, usual-paced.
aPearson product-moment correlation (all other: Spearman rank-order correlation).
*, Correlation is significant at the 5% level (2-tailed). Bold formatting indicates correlations ≥0.4 (moderate) or ≥0.6 (strong).
**, Correlation is significant at the 1% level (2-tailed).
Associations between clinical tests and falls history are reported in Supplementary Table 3. People with HIV reporting FOF had poorer scores in total PPB (P = .008), PPB narrow walk (P = .048), and PPB chair rise (P = .008). Those with FOF also performed more poorly in SLS Eyes Closed (P = .004), had slower 5STS times (P = .003), and fewer 30sSTS repetitions (P = .021). The PPB balance subscore was the only test with a significant association with recent falls (P < .001). The EGVI scores were not correlated to HIV duration or ART use (Supplementary Table 4).
DISCUSSION
This study is the first to describe biomechanical differences in the gait of PWH relative to HIV-free peers. Our findings further suggest the 5STS as a clinical test that might be relevant for early screening for gait deviations in PWH, given the correlations with increased gait variability and self-reported functional problems. Young-to-middle-aged PWH without obvious predisposing factors for mobility problems presented with a slowed and more variable gait pattern. The observed deviations were subtle, and had we included PWH with peripheral neuropathy, gait differences might perhaps have been even more pronounced. Chair stand performance was slower in PWH, and slowed 5STS performance in PWH was related to increased gait variability, worse self-reported function, and FOF. People with HIV walked with a significantly more variable gait than SNP, which may be interpreted as unstable walking [9, 10]. Adopting a more variable gait may suggest impaired motor control (particularly implicating brain areas important for sensorimotor integration and coordination [40]). The associated foot placement errors and instability may be compounded in PWH by prolonged automated postural response latencies and abnormal postural reflex regulation [2]. People with HIV may thus have an increased falls risk when exposed to unexpected perturbations. Although there was no correlation between EGVI scores and recent falls events, the value of the EGVI in predicting falls warrants further investigations.
People with HIV gait demonstrated significant slowing, evidenced by increased time-related parameters such as absolute time spent in stance and producing steps. Because this remained evident after controlling for covariables, HIV-related disease physiology may be indicated as a contributing factor (eg, white matter alterations [41]). Our study provides new information about speed impairments. A systematic review suggested that PWH demonstrate slowed speeds when having to walk fast, but not when walking at a usual pace [2]. However, most of the studies reporting differences used predicted norms from the literature for comparison. Because gait patterns strongly relate to context [42], between-group differences may not be reliable when comparing data from different populations. Using an HIV-free comparison group from the same community, we found that when walking fast, PWH overcame impairments noted for usual-paced or DT gait. However, we did not determine whether this required more effort (eg, whether maintaining such a pace for an extended period would be possible without fatigue). Further research should investigate how long such performance can be maintained, and whether falls risk increases due to PWH exceeding their capacity for safe walking. The possibility that PWH can tap into resources to walk fast with kinematics similar to SNP may imply that early gait impairments may be reversible; however, this requires further research.
Under DT conditions, PWH walked slower with an increased step time, but also with longer stride lengths. The longer strides in PWH may perhaps be explained by prioritization of the counting task over walking. This resembles an adaptation observed in older adults during dual tasking [44]. In contrast, in young healthy adults, stride characteristics usually remain unchanged because the DT cost does not represent a high cognitive load. Dual-task walking additionally revealed deviations in PWH’s joint ROM in a kinematic pattern resembling the age-related biomechanical plasticity (ie, increased hip and decreased distal mechanical output when walking) often seen in older adults [24]. Biomechanical plasticity is likely related to changes in neural and musculoskeletal function [44], and the associated walking strategy involves a distal-to-proximal shift in muscle function and thus joint ROM [24]. The underlying mechanism in PWH may perhaps be similar, given the potential involvement of central neuromotor control [45].
The significantly poorer 5STS performance by PHW concurs with studies of middle-aged PWH (~40–50 years) in high-income settings [4, 46]. Dose-response relationships exist between (1) slowed 5STS and mortality [4] and (2) good performance and greater QOL [3]. The clinical importance of the 5STS is further supported by studies showing that half of PWH in their forties demonstrate impaired performance, one third may further decline over time, and a poor baseline score (average: 9.6 seconds) predicts falls [46].
To understand the moderate correlations between 5STS and 3DGA, the aspects of function that they measure must be considered. Gait biomechanics may be categorized under the body function domain on the International Classification of Functioning, Disability and Health [47], whereas the activity domain better represents the chair-rise movement. However, body functions are inherently related to functional activities. It has been demonstrated that rapid force-relaxation, rather than muscle activation, is affected in the knee extensors of cART-managed PWH [48]. Studies in other populations have also suggested that impaired muscle power (more so than weakness) can influence functional mobility [49]. Although our 3DGA did not include kinetics to quantify muscle power, a hypothesis of impaired lower limb muscle power in PWH would provide a feasible link between the observations for 5STS performance and gait biomechanics. The increased variability in PWH’s gait may indicate reduced dynamic stability and less accurate responses to perturbations. The 5STS requires rapid and coordinated lower limb biomechanics and dynamic balance, thus one might expect to observe at least some correlation with gait variability. However, 5STS also represents a different movement sequence than that involved in walking, and both activities may be influenced by factors not measured in this study.
Slowed 5STS also significantly correlated with poorer self-reported function and FOF. Although not yet studied in PWH, FOF is a major and complex health problem in community-dwelling elderly and may present in those with and without previous falls [50]. Clinical tests associated with FOF may already be able to identify potential falls risk in PWH; however, sensitive falls-risk screening tests should be investigated in prospective longitudinal research.
Limitations
The cross-sectional design cannot infer causality. The study provides insights into young-to-middle-aged South African PWH without peripheral neuropathy or obesity, drawn from 2 clinical settings, compared with community-matched SNP. The results are not necessarily generalizable to other populations. Although we excluded individuals with pathologies such as neuropathy and injuries that may affect gait, we did not exclude other comorbidities that can affect gait, such as chronic pain. The magnitude of angular differences between-groups was small. However, the differences all exceeded measurement errors of the 3DGA system [18]. The EGVI calculation requires a minimum of 5 absolute differences [8, 25], which necessitated the exclusion of some datasets. Furthermore, this study used a limited number of steps, ie, the absolute minimum required for calculating the EGVI, due to the available space for 3DGA in the clinical settings. More steps will likely provide more robust results.
CONCLUSIONS
In this South African rural sample, relatively young PWH presented with biomechanical walking deviations suggestive of a slowed and unstable gait pattern, especially under cognitive challenges. Apart from improving understanding of gait patterns in PWH, findings also underscore the need for early screening of PWH for motor impairments. Five-times STS may be useful to predict gait deviations in PWH. Further research is needed to determine whether the 5STS test predicts falls and whether the impairments noted in PWH are reversible.
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Acknowledgments
We thank study participants for their time and effort and the Cape Winelands HAART to HEART team, research nurses, and human immunodeficiency virus counselors for assisting with data collection and study logistics. In addition, we thank K. Grimmer for providing writing support for this manuscript. We gratefully acknowledge the relevant Community Health Clinics and Stellenbosch University Ukwanda Centre for Rural Health for hosting data collection.
Author contributions. K. B. and Q. L. contributed to conceptualization; K. B., Q. L., and A. G. contributed to methodology; K. B. contributed to formal analysis; K. B., Q. L., H. S., M. F. E., and I. W. contributed to investigation; Q. L., H. S., M. F. E., I. W., and K. B. contributed to resources; K. B. contributed to writing the original draft; K. B., Q. L., A. G., H. S., M. F. E., and I. W. contributed to writing, review, and editing; K. B., A. G., H. S., M. F. E., I. W., and Q. L. were all involved in data interpretation, manuscript writing, revision, and critical comment; all authors approved the final version for publication.
Disclaimer. The views and opinions expressed are not those of the funders but of the authors of the material publicized.
Financial support. This research was funded by the South African Medical Research Council under a Self-Initiated Research Grant (to Q. L.) and the Harry Crossley Foundation (Grants SU-PT-15/10-000005 and SU-PT-16/10-000057; to K. B.). The PhD from which this study emanated was funded by the South African Medical Research Council in terms of the National Health Scholars Programme (to K. B.) from funds provided for this purpose by the National Department of Health. K. B. was additionally funded by Subcommittee C of the Research Committee’s Early-Career Research Fund, Stellenbosch University.
Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.
References
- 1.Quigley A, MacKay-Lyons M. Physical deficits among people living with HIV: a review of the literature and implications for rehabilitation. Phys Ther Rev 2020; 25:29–41. [Google Scholar]
- 2.Berner K, Morris L, Baumeister J, Louw Q. Objective impairments of gait and balance in adults living with HIV-1 infection: a systematic review and meta-analysis of observational studies. BMC Musculoskelet Disord 2017; 18:325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Erlandson KM, Allshouse AA, Jankowski CM, et al. Relationship of physical function and quality of life among persons aging with HIV infection. AIDS 2014; 28:1939–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Greene M, Covinsky K, Astemborski J, et al. The relationship of physical performance with HIV disease and mortality. AIDS 2014; 28:2711–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Beauchet O, Annweiler C, Callisaya ML, et al. Poor gait performance and prediction of dementia: results from a meta-analysis. J Am Med Dir Assoc 2016; 17:482–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.O’Sullivan M, Blake C, Cunningham C, et al. Correlation of accelerometry with clinical balance tests in older fallers and non-fallers. Age Ageing 2009; 38:308–13. [DOI] [PubMed] [Google Scholar]
- 7.Roiz dMR, Cacho EWA, Pazinatto MM, et al. Gait analysis comparing Parkinson’s disease with healthy elderly subjects. Arq Neuropsiquiatr 2010; 68:81–86. [DOI] [PubMed] [Google Scholar]
- 8.Gouelle A, Mégrot F, Presedo A, et al. The gait variability index: a new way to quantify fluctuation magnitude of spatiotemporal parameters during gait. Gait Posture 2013; 38:461–5. [DOI] [PubMed] [Google Scholar]
- 9.Hausdorff JM, Rios DA, Edelberg HK. Gait variability and fall risk in community-living older adults: a 1-year prospective study. Arch Phys Med Rehabil 2001; 82:1050–6. [DOI] [PubMed] [Google Scholar]
- 10.Brach JS, Perera S, Studenski S, et al. Meaningful change in measures of gait variability in older adults. Gait Posture 2010; 31:175–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Berner K, Strijdom H, Essop MF, et al. Fall history and associated factors among adults living with HIV-1 in the Cape Winelands, South Africa: an exploratory investigation. Open Forum Infect Dis 2019; 6:ofz401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bauer LO, Wu Z, Wolfson LI. An obese body mass increases the adverse effects of HIV/AIDS on balance and gait. Phys Ther 2011; 91:1063–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sharma A, Hoover DR, Shi Q, et al. Longitudinal study of falls among HIV-infected and uninfected women: the role of cognition. Antivir Ther 2018; 23:179–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Erlandson KKM, Allshouse AAAA, Jankowski CCM, et al. Risk factors for falls in HIV-infected persons. J Acquir Immune Defic Syndr 2012; 61:484–489. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Charumbira MY, Berner K, Louw QA. Falls in people living with HIV: a scoping review. BMJ Open 2020; 10:e034872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Peters A, Galna B, Sangeux M, et al. Quantification of soft tissue artifact in lower limb human motion analysis: a systematic review. Gait Posture 2010; 31:1–8. [DOI] [PubMed] [Google Scholar]
- 17.McGinley JL, Baker R, Wolfe R, Morris ME. The reliability of three-dimensional kinematic gait measurements: a systematic review. Gait Posture 2009; 29:360–9. [DOI] [PubMed] [Google Scholar]
- 18.Berner K, Cockcroft J, Louw Q. Kinematics and temporospatial parameters during gait from inertial motion capture in adults with and without HIV: a validity and reliability study. Biomed Eng Online 2020; 19:57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lee LW, Zavarei K, Evans J, et al. Reduced hip extension in the elderly: dynamic or postural? Arch Phys Med Rehabil 2005; 86:1851–4. [DOI] [PubMed] [Google Scholar]
- 20.Kerrigan DC, Lee LW, Collins JJ, et al. Reduced hip extension during walking: healthy elderly and fallers versus young adults. Arch Phys Med Rehabil 2001; 82:26–30. [DOI] [PubMed] [Google Scholar]
- 21.Kemoun G, Thoumie P, Boisson D, Guieu JD. Ankle dorsiflexion delay can predict falls in the elderly. J Rehabil Med 2002; 34:278–83. [DOI] [PubMed] [Google Scholar]
- 22.Boyer KA, Johnson RT, Banks JJ, et al. Systematic review and meta-analysis of gait mechanics in young and older adults. Exp Gerontol 2017; 95:63–70. [DOI] [PubMed] [Google Scholar]
- 23.Mortaza N, Abu Osman NA, Mehdikhani N. Are the spatio-temporal parameters of gait capable of distinguishing a faller from a non-faller elderly? Eur J Phys Rehabil Med 2014; 50:677–91. [PubMed] [Google Scholar]
- 24.Beijersbergen CM, Granacher U, Vandervoort AA, et al. The biomechanical mechanism of how strength and power training improves walking speed in old adults remains unknown. Ageing Res Rev 2013; 12:618–27. [DOI] [PubMed] [Google Scholar]
- 25.Gouelle A, Rennie L, Clark DJ, et al. Addressing limitations of the Gait Variability Index to enhance its applicability: the enhanced GVI (EGVI). PLoS One 2018; 13:e0198267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Balasubramanian CK, Clark DJ, Gouelle A. Validity of the gait variability index in older adults: effect of aging and mobility impairments. Gait Posture 2015; 41:941–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Swanenburg J, de Bruin ED, Favero K, et al. The reliability of postural balance measures in single and dual tasking in elderly fallers and non-fallers. BMC Musculoskelet Disord 2008; 9:162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Studenski S, Perera S, Wallace D, et al. Physical performance measures in the clinical setting. J Am Geriatr Soc 2003; 51:314–22. [DOI] [PubMed] [Google Scholar]
- 29.Bohannon RW. Single limb stance times—A descriptive meta-analysis of data from individuals at least 60 years of age. Top Geriatr Rehabil 2006; 22:70–77. [Google Scholar]
- 30.Simonsick EM, Newman AB, Nevitt MC, et al. ; Health ABC Study Group. Measuring higher level physical function in well-functioning older adults: expanding familiar approaches in the Health ABC study. J Gerontol A Biol Sci Med Sci 2001; 56:M644–9. [DOI] [PubMed] [Google Scholar]
- 31.Louwagie GM, Bachmann MO, Meyer K, et al. Highly active antiretroviral treatment and health related quality of life in South African adults with human immunodeficiency virus infection: a cross-sectional analytical study. BMC Public Health 2007; 7:244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Delate T, Coons SJ. The use of 2 health-related quality-of-life measures in a sample of persons infected with human immunodeficiency virus. Clin Infect Dis 2001; 32:E47–52. [DOI] [PubMed] [Google Scholar]
- 33.Mathews WC, May S. EuroQol (EQ-5D) measure of quality of life predicts mortality, emergency department utilization, and hospital discharge rates in HIV-infected adults under care. Health Qual Life Outcomes 2007; 5:5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Rudy BJ, Murphy DA, Harris DR, et al. ; Adolescent Trials Network for HIV/AIDS Interventions. Patient-related risks for nonadherence to antiretroviral therapy among HIV-infected youth in the United States: a study of prevalence and interactions. AIDS Patient Care STDS 2009; 23:185–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wang C-Y,, Olson, SL,, Protas, EJ,Test-retest strength reliability: hand-held dynamometry in community-dwelling elderly fallers. Arch Phys Med Rehabil 2002; 83:811–5. [DOI] [PubMed] [Google Scholar]
- 36.Kressig RW, Beauchet O; European GAITRite Network Group. Guidelines for clinical applications of spatio-temporal gait analysis in older adults. Aging Clin Exp Res 2006; 18:174–6. [DOI] [PubMed] [Google Scholar]
- 37.Lim CR, Harris K, Dawson J, et al. Floor and ceiling effects in the OHS: an analysis of the NHS PROMs data set. BMJ Open 2015; 5:e007765. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Beck CW, Bliwise NG. Interactions are critical. CBE Life Sci Educ 2014; 13:371–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Mukaka MM. Statistics corner: a guide to appropriate use of correlation coefficient in medical research. Malawi Med J 2012; 24:69–71. [PMC free article] [PubMed] [Google Scholar]
- 40.Tian Q, Chastan N, Bair WN, et al. The brain map of gait variability in aging, cognitive impairment and dementia-A systematic review. Neurosci Biobehav Rev 2017; 74:149–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Sullivan EV, Rosenbloom MJ, Rohlfing T, et al. Pontocerebellar contribution to postural instability and psychomotor slowing in HIV infection without dementia. Brain Imaging Behav 2011; 5:12–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Ebersbach G, Sojer M, Müller J, et al. Sociocultural differences in gait. Mov Disord 2000; 15:1145–7. [DOI] [PubMed] [Google Scholar]
- 43.Li KZ, Abbud GA, Fraser SA, Demont RG. Successful adaptation of gait in healthy older adults during dual-task treadmill walking. Neuropsychol Dev Cogn B Aging Neuropsychol Cogn 2012; 19:150–67. [DOI] [PubMed] [Google Scholar]
- 44.Hortobágyi T, Rider P, Gruber AH, DeVita P. Age and muscle strength mediate the age-related biomechanical plasticity of gait. Eur J Appl Physiol 2016; 116:805–14. [DOI] [PubMed] [Google Scholar]
- 45.Scott WB, Oursler KK, Katzel LI, et al. Central activation, muscle performance, and physical function in men infected with human immunodeficiency virus. Muscle Nerve 2007; 36:374–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Richert L, Brault M, Mercié P, et al. ; Groupe d’Epidémiologie Clinique du SIDA en Aquitaine (GECSA). Decline in locomotor functions over time in HIV-infected patients. AIDS 2014; 28:1441–9. [DOI] [PubMed] [Google Scholar]
- 47.World Health Organisation. International Classification of Functioning, Disability and Health: ICF. Geneva: World Health Organization; 2001. [Google Scholar]
- 48.Russ DW, Scott WB, Oursler KK, King JS. Paradoxical contractile properties in the knee extensors of HIV-infected men treated with antiretroviral therapy. Appl Physiol Nutr Metab 2010; 35:713–7. [DOI] [PubMed] [Google Scholar]
- 49.Bean JF, Leveille SG, Kiely DK, et al. A comparison of leg power and leg strength within the InCHIANTI study: which influences mobility more? J Gerontol A Biol Sci Med Sci 2003; 58:728–33. [DOI] [PubMed] [Google Scholar]
- 50.Scheffer AC, Schuurmans MJ, van Dijk N, et al. Fear of falling: measurement strategy, prevalence, risk factors and consequences among older persons. Age Ageing 2008; 37:19–24. [DOI] [PubMed] [Google Scholar]
- 51.Bennett JA, Winters-Stone K, Nail LM, Scherer J. Definitions of sedentary in physical-activity-intervention trials: a summary of the literature. J Aging Phys Act 2006; 14:456–77. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.